JPH0668241A - Fingerprint identifying device - Google Patents

Abstract

PURPOSE:To provide a fingerprint identifying device in which the generation of an erroneous identification can be reduced. CONSTITUTION:Fingerprint picture data inputted from a fingerprint inputting device 1 are scanned by a matching degree detecting device 3 based on a registered partial picture stored in a registered picture storage device 6, and the unmatching degree is searched. The minimum unmatching degree obtained by the result is added to a unmatching degree adding memory 4. Then, the added result is checked by a person himself deciding device 5, and when the value is beyond a threshold value, the identified result is decided to be an another person, and when the value is not beyond the threshold value, the similar processing is repeated based on the next registered partial picture. Then, when the value of the unmatching degree adding memory 4 is not beyond the prescribed threshold value after the processing is repeated to the entire registered partial pictures, the identified result is decided to be the person himself.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a fingerprint identification device,
In particular, the present invention relates to a fingerprint identification device that performs identification based on an input fingerprint image and a registered partial image.

[0002]

2. Description of the Related Art There is known a method of inputting a fingerprint image and identifying an individual based on the degree of coincidence with a partial image registered in advance.

For example, it is assumed that an identification process is performed on an input fingerprint image 41 as shown in FIG. 6A obtained by reading a fingerprint. In this case, scanning is performed within predetermined areas in the input fingerprint image based on the four registered partial images T _A , T _B , T _C, and T _D as shown in FIG. 6B. Find the degree of disagreement between images. Here, the degree of disagreement is the proportion of pixels whose values do not match when the binarized images are overlapped, and takes a value of 0 if they completely match and 100 if they do not match at all. As a result, the minimum degree of disagreement in the scanning area and the position of each of the partial images T _{A to} T _D are determined as shown in FIG. 6C. The content of the fingerprint identification method using this minimum degree of disagreement is shown in the flowchart of FIG.

First, in step S41, the value of the total number stored in the memory is set to 0, and the fingerprint is read, that is, the image is input (S42). Then, in step S43, scanning is performed within the scanning area of the input fingerprint image based on the partial image data, and the degree of mismatch is calculated. From the result of the calculation of the degree of disagreement within the scanning region, the minimum degree of disagreement is detected in step S44. Then step S
At 45, it is judged if the detected minimum disagreement degree is less than or equal to a predetermined threshold value. If the minimum dissimilarity is equal to or less than the threshold value, in step S46, 1 is added to the total number stored in the memory, and step S47 is performed.
Then, it is determined whether or not the value of the total number is equal to or larger than the reference value. When the value of the total number is equal to or larger than the reference value, it is determined in step S49 that the read fingerprint matches the preregistered fingerprint, that is, the identification result is the original person, and the process ends.

On the other hand, in step S45, if the minimum degree of disagreement is not less than the threshold value, steps S46 and S4.
Skip 7 and return to step S43. The processes of steps S43 to S47 are repeated for the number of registered partial images. Then, as a result of repeating the processing by the number of registered partial images, if the value of the total number does not exceed the reference value in step S47, the flow proceeds to step S48, and the read fingerprint is registered in advance. If the fingerprint does not match the fingerprint, that is, the identification result is determined to be another person, the flow ends.

[0006]

In the conventional fingerprint identification method as described above, it is determined for each registered partial image whether the degree of mismatch is less than or equal to the threshold value. Therefore, even if the fingerprint to be identified and the registered fingerprint are the same person, due to the effects of noise and distortion of the read image data at the time of input, the degree of inconsistency of some partial images will be slightly If it exceeds, even if the degree of non-coincidence of the other partial images is very small, there is a risk of being erroneously determined to be another person. To prevent such erroneous detection, if the threshold for determining the degree of mismatch is increased,
On the contrary, the possibility that another person will be mistakenly judged as the true person increases.

The present invention has been made to solve the above problems, and an object of the present invention is to provide a fingerprint identification device in which erroneous identification is less likely to occur.

[0008]

According to another aspect of the present invention, there is provided a fingerprint identification device which reads a fingerprint to be identified and inputs fingerprint image data, and a fingerprint image obtained by reading the identified fingerprint. A storage unit that stores a plurality of partial image data forming data and a mismatch detection unit that compares the input fingerprint image data and each of the partial image data to detect each mismatch, and the detected mismatch And a determination unit that determines that the fingerprint to be identified and the identified fingerprint belong to the same person when the total value of the degrees is less than or equal to a predetermined value.

According to a second aspect of the present invention, there is provided a fingerprint identification device for reading fingerprints of a plurality of fingers to be identified and inputting fingerprint image data for each finger, and inputting means for each of a plurality of identified fingers of the same person. Storage means for storing a plurality of partial image data constituting each of the fingerprint image data obtained by reading the fingerprint, and stored for each specified finger corresponding to the fingerprint image data input for each finger to be identified. And the dissimilarity detection means for detecting the dissimilarity of each of the partial image data for each finger, and the total value of the dissimilarity detected for any one finger is the first predetermined value. When it is less than or equal to the value, it is determined that the fingerprint of the finger to be identified and the fingerprint of the specified finger belong to the same person, and when the total value for any one finger exceeds the first predetermined value, a plurality of Each degree of disagreement detected for each finger of When each of the total value of the value is less than the second predetermined value, the fingerprint is identified as the fingerprint of the finger to be identified finger is that a the determining means is of the same person.

[0010]

According to the first aspect of the invention, the fingerprint is identified based on the total value of the detected inconsistencies.

According to the second aspect of the present invention, when the fingerprint cannot be identified by the total value for one finger, the fingerprint is identified based on the total value for each finger.

[0012]

1 is a block diagram showing the configuration of a fingerprint identification device according to a first embodiment of the present invention.

Referring to the drawing, the fingerprint image data input from the fingerprint input device 1 is binarized after preprocessing such as noise removal is performed by the preprocessing device 2. Obtained 2
With respect to the binarized input image data, for example, by the coincidence detection device 3 configured by a microcomputer,
For example, scanning is performed on the basis of the partial image data stored in the registered image storage device formed of an EEPROM, a CDROM or the like, and the degree of inconsistency of mutual data is obtained. The minimum disagreement degree thus obtained is added to the disagreement degree addition memory 4 configured by, for example, a DRAM or the like, and the identity determination device 5 configured by a microcomputer, for example, is used.
The contents of can be examined. Then, if the checked memory content exceeds a predetermined threshold value, the identification result is determined to be another person, and the process ends. On the other hand, when the examined memory contents do not exceed the predetermined threshold value, the matching degree detection device 3 repeats the same processing as described above for the next registered partial image. After repeating the same processing for all registered partial images, if the content of the mismatch degree addition memory does not exceed the predetermined threshold value, the identification result is determined to be the person, and the identification processing is performed. Ends.

FIG. 2 is a flow chart showing specific identification processing contents of the fingerprint identification apparatus shown in the embodiment of FIG.

First, in step S1, the value of the mismatch degree addition memory is set to 0 as an initial setting, and then a fingerprint is read, that is, image data is input (S2). Then, based on the registered partial image data, the input fingerprint image is scanned in a predetermined area, and the degree of mismatch between the data is calculated (S3). Then, in step S4, the minimum degree of non-coincidence within the scanning area is obtained, and the obtained minimum degree of non-coincidence is added to the discrepancy degree addition memory (S5).

Next, in step S6, it is determined whether or not the value of the mismatch degree addition memory exceeds a predetermined threshold value. If the memory value exceeds the predetermined threshold value, it is determined in step S7 that the identification result is another person, and the identification process ends.

On the other hand, in step S6, if the value of the mismatch degree addition memory is less than or equal to the threshold value, step S6.
The processing from 3 to step S6 is repeated for the number of registered partial images. Then, when the above process is repeated for all partial images, and the value of the mismatch degree addition memory is equal to or less than a predetermined threshold value,
The flow proceeds to step S8, the identification result is determined to be the person, and the identification process ends.

In this embodiment, the registered partial image
, As shown in FIG. 6 of the conventional example, T _A~ T_DThe four parts of
A partial image T_ADisagreement of m_AAnd the department
Minute image T_BDisagreement of m_BAnd the partial image T_CDisagreement
Degree in m_CAnd, and the partial image T_DDisagreement of m_DTosu
Then, the judgment condition of step S6 is expressed by the following equation.
To be done.

_{M A} + m _B + m _C + m _D ≦ threshold It should be noted that in the above embodiment, the weights of the degree of disagreement with respect to each registered partial image are all assumed to be equal to 1, as is clear from the above equation. . However, it is also possible to add the respective dissimilarities by weighting the above formula according to the reliability of the dissimilarities with respect to each partial image. FIG. 3 is a diagram showing the positional relationship between the input fingerprint image shown as an example of this weighting and the registered partial image.

In the figure, "X" indicates a reference point that serves as a reference for the original position of each registered partial image, where A is the partial image T _A , B is the partial image T _B , and C is D of the partial image T _C indicates the position (shown by a broken line) in which the partial image T _D should originally be. However, in reality, the position shown by the solid line in the figure, that is, A ',
It is assumed that the positions of B ′, C ′ and D ′ are the positions where the degree of mismatch between the partial images is the minimum. From this result, the values of the distances d _A to d _D indicate the displacement of each partial image from the original position to the position where the degree of disagreement is minimized. Then, by defining the weighting function w (d) with the value of this positional deviation as a reference, the reliability of the above equation can be further improved. That is, this weighting function may be weighted such that the smaller the value of the positional deviation, the better the evaluation value of the identification result. This is because, if the image data is for another person, there is no specific relationship between the position where the partial image should be and the position where the degree of disagreement is minimal, so the position shift value is generally large. Because. By doing this, the evaluation value becomes worse between other people, and it is possible to further eliminate erroneous identification with other people. The judgment conditions to which the above weighting is added are expressed as follows.

W (d _A ) m _A + w (d _B ) m _B + w (d
_C ) mc + w (d _D ) m _D <threshold value As described above, the sum of the degree of disagreement for each partial image is used for the determination of discrimination. Generally, the degree of disagreement takes a large value between data of other persons. However, it is easier to separate the relationship between other people and the relationship between other people because it is distributed in a small value among the data of the other people, and as a result, while eliminating the possibility of false detection of other people, This is because the identification rate can be improved.

FIG. 4 is a block diagram showing the arrangement of a fingerprint identifying apparatus according to the second embodiment of the present invention.

In the following description, differences from the device of FIG. 1 showing the first embodiment will be mainly described.

Referring to the drawing, the multi-fingerprint input device 21
Input image data from fingerprints of multiple fingers. The mismatch degree addition memories 24a, 24b, ... Are prepared for the number of fingers to be input, and the registered image storage device 26 also stores partial images registered in advance corresponding to the number of fingers to be input. A plurality of fingers are stored. As the fingerprint identification process, the same process as the fingerprint identification process shown in the first embodiment is performed on the input fingerprint of each finger.

FIG. 5 is a flow chart showing the contents of the fingerprint identification processing according to the second embodiment of the present invention.

In the present embodiment, for the sake of convenience of explanation, an example is shown in which the number of fingers to be input is two.

First, in step S31, the values of the mismatch degree addition memory 1 and the mismatch degree addition memory 2 are cleared to 0 as an initial setting. Next, in step S32, the image data of each of the two fingers is input, and the degree of mismatch of the partial images is calculated for each finger. That is, step S
In 33, first, the partial image registered in advance for the finger input first is scanned within the range of the predetermined area, and the degree of mismatch is calculated. Then, in step S34, the minimum degree of mismatch is detected from the result of the degree of mismatch calculation, and the value is added to the mismatch degree addition memory 1 (S35). The processes of steps S33 to S35 are repeated for the number of partial images corresponding to the first finger. Then, in step S36, it is determined whether or not the value of the mismatch degree addition memory 1 is less than or equal to the threshold value set for one finger.
If the result of the determination is that the value in the mismatch degree addition memory 1 is less than or equal to the threshold value for one finger, the flow proceeds to step S37, the identification result is determined to be the person, and the subsequent processing ends.

On the other hand, if the value of the mismatch degree addition memory 1 is not less than the threshold value for one finger in step S36, the same process as described above is performed for the second finger. Also for the second finger, step S36.
If the value of the mismatch degree addition memory 2 is less than or equal to the threshold value of one finger as a result of the determination of No. 1, the flow also follows Step S3.
In step 7, the identification result is determined to be the person, and the process ends. As a result of the processing performed on the second finger, if the value in the mismatch degree addition memory 2 is not less than or equal to the threshold value set for the one finger, the flow proceeds to step S38, and the mismatch degree addition is performed. The values of the memory 1 and the mismatch degree addition memory 2 are summed. It is determined whether or not the total value thereof is less than or equal to the threshold value set for two fingers. When the total value is less than or equal to the threshold value set for two fingers, the flow proceeds to step S37, the identification result is determined to be the person, and the process ends. On the other hand, if the total value is not less than or equal to the threshold value for two fingers, the flow is step S39.
Then, the identification result is judged to be another person, and the process ends.

The above-described identification process is performed on a plurality of fingers. Generally, there are few cases where the fingerprints of two fingers are similar to each other among other people. This is because when the result is not clear, the possibility of being erroneously detected as another person is further eliminated by using the total of the disagreement degrees of the entire two fingers. By performing the identification process using a plurality of fingers in this way, even if the input state of the fingerprint image of one finger is a little bad, it is determined as the plurality of fingers as a whole, and thus the identification of the person is made. It is possible to further improve the rate. Further, as described above, even when a plurality of fingers are read, if one of the fingers satisfies the identification condition that the person is the original person, the identification result can be determined to be the original person. Even when the finger is scratched or the input state of only one finger is particularly bad, it can be sufficiently dealt with, and the identification rate of the person is further improved.

Even when a plurality of fingers are input to identify them, the degree of non-coincidence of each partial image is weighted as in the one-finger identifying method of the first embodiment. May be added.

Further, in the above-mentioned second embodiment, the input of two fingers is taken as an example, but it goes without saying that the present invention can be similarly applied to the input of three or more fingers.

[0032]

As described above, the invention of claim 1 is as follows.
Since the fingerprint is identified based on the total value of the detected inconsistencies, the fingerprint identification process can be performed with high accuracy.

The invention of claim 2 is, as explained above, 1
When the fingerprint cannot be identified by the total value of the degree of disagreement with respect to the fingers of the book, the fingerprint is identified based on the total value for each finger, so that more accurate fingerprint identification processing can be performed.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a fingerprint identification device according to a first embodiment of the present invention.

FIG. 2 is a flowchart showing the contents of processing performed by the fingerprint identification device of FIG.

FIG. 3 is a diagram for explaining the content of weighting when weighting processing is performed as a modification of the processing in the first embodiment of the present invention.

FIG. 4 is a block diagram showing a configuration of a fingerprint identification device according to a second embodiment of the present invention.

5 is a flowchart showing the processing contents of the fingerprint identification device of FIG.

FIG. 6 is a diagram for explaining the content of the identification method of the conventional fingerprint identification device.

FIG. 7 is a flowchart showing the processing contents of a conventional fingerprint identification device.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Fingerprint input device 3 Matching degree detection device 4 Mismatch degree addition memory 5 Identity determination device 6 Registered image storage device 21 Multiple fingerprint input device 23 Matching degree detection device 24a, 24b Mismatch degree addition memory 25 Principal determination device 26 Registration image storage device

Claims (2)

[Claims] 1. An input unit for reading a fingerprint to be identified and inputting fingerprint image data, and a storage unit for storing a plurality of partial image data constituting the fingerprint image data obtained by reading the identified fingerprint. Comparing each of the input fingerprint image data and each of the partial image data, a non-coincidence degree detecting unit for detecting each non-coincidence degree, when the total value of each of the detected non-coincidence degrees is a predetermined value or less, A fingerprint identification device comprising: a determination unit that determines that the fingerprint to be identified and the identified fingerprint belong to the same person. 2. An input unit for reading fingerprints of a plurality of fingers to be identified and inputting fingerprint image data for each finger, and a fingerprint image obtained by reading fingerprints for a plurality of identified fingers of the same person. Storage means for storing a plurality of partial image data constituting each of the data, the fingerprint image data input for each finger to be identified,
Dissimilarity detection means for comparing each of the corresponding partial image data stored for each of the specified fingers for each finger, and detecting the dissimilarity of each, the detection being performed for any one finger. When the total value of the disagreement degrees is less than or equal to a first predetermined value, it is determined that the fingerprint of the finger to be identified and the fingerprint of the identified finger belong to the same person, and any one of the above When the total value for each finger exceeds the first predetermined value, and when each total value of the total values of the detected dissimilarity degrees for each of the plurality of fingers is equal to or less than a second predetermined value, the identification is performed. A fingerprint identification device, comprising: a determination unit that determines that the fingerprint of a finger to be printed and the fingerprint of the specified finger belong to the same person.